movie_searcher 0.0.6 → 0.0.8

data/Gemfile.lock

@@ -1,7 +1,7 @@
 PATH
   remote: .
   specs:
-    movie_searcher (0.0.6)
+    movie_searcher (0.0.7)
       hintable_levenshtein
       httparty
 
@@ -38,5 +38,7 @@ DEPENDENCIES
   autotest-fsevent
   autotest-growl
   autotest-standalone
+  hintable_levenshtein
+  httparty
   movie_searcher!
   rspec

data/lib/imdb_party/exclude.yaml

@@ -0,0 +1,22 @@
+excluded:
+  - EXTENDED
+  - "((xvid(-[a-z]+)?))"
+  - "(720|480)p"
+  - B(R|D)Rip
+  - cam
+  - telesync|([^a-z])ts([^a-z]?)
+  - telecine|([^a-z])tc([^a-z]?)
+  - "([^a-z])rx([^a-z]?)"
+  - dvdscr
+  - screener
+  - hdtv
+  - dvdrip
+  - workprint
+  - brrip|bluray|blu(-?)ray
+  - dvd(-?)r
+  - x264
+  - UNRATED
+  - IMAGiNE
+  - SCR
+  - "((ac3(-[a-z]+)?))"
+  - LIMITED

data/lib/imdb_party/movie.rb

@@ -1,6 +1,6 @@
 module ImdbParty
   class Movie
-    attr_accessor :imdb_id, :title, :directors, :writers, :tagline, :company, :plot, :runtime, :rating, :poster_url, :release_date, :certification, :genres, :actors, :trailers
+    attr_accessor :imdb_id, :title, :directors, :writers, :tagline, :company, :plot, :runtime, :rating, :poster_url, :release_date, :certification, :genres, :actors, :trailers, :year
 
     def initialize(options={})
       if not options.nil? and options.keys.first.class == Symbol

data/lib/movie_searcher.rb

@@ -2,7 +2,7 @@ require "imdb_party"
 require 'imdb_party/levenshtein'
 
 class MovieSearcher
-  attr_accessor :options
+  attr_accessor :options, :cleaners
   
   def initialize(args)
     args.keys.each { |name| instance_variable_set "@" + name.to_s, args[name] unless name == :options }
@@ -15,6 +15,8 @@ class MovieSearcher
     }
     
     @options.merge!(args[:options]) unless args[:options].nil?
+    
+    @cleaners = YAML.load(File.read("#{File.dirname(__FILE__)}/imdb_party/exclude.yaml"))["excluded"]
   end
   
   def self.find_by_release_name(search_value, options = {})
@@ -29,7 +31,7 @@ class MovieSearcher
   end
   
   def to_long?
-    @split = @search_value.split(@options[:split])
+    @split = self.cleaner(@search_value).split(@options[:split])    
     @split.length > @options[:long]
   end
   
@@ -39,17 +41,14 @@ class MovieSearcher
     until current <= 0 do
       title = @split.take(current).join(' ')
       movies = @options[:imdb].find_by_title(title)
-      break if movies.any?
+      break if movies.any? and movies.reject{ |movie| self.shortest(movie, title).nil? }.any?
       current -= 1 
     end
     
     return if movies.nil? or not movies.any?
-    
-    # Cleaning up the title, no years allowed
-    title = title.gsub(/[^a-z0-9]/i, '').gsub(/(19|20)\d{2}/, '')
-    
+
     movie = movies.map do |movie| 
-      [movie, Levenshtein.distance(movie[:title].gsub(/[^a-z0-9]/i, ''), title, @options[:limit])]
+      [movie, self.shortest(movie, title)]
     end.reject do |value|
       value.last.nil?
     end.sort_by do |_,value|
@@ -58,12 +57,36 @@ class MovieSearcher
     
     return if movie.nil?
     
-    return ImdbParty::Movie.new(movie.first)
+    ImdbParty::Movie.new(movie.first)
   end
   
   def self.method_missing(method, *args, &block)  
     result = ImdbParty::Imdb.new.send(method, *args)
-    return if result.nil?
     result.class == Array ? result.map{|r| ImdbParty::Movie.new(r)} : result
   end
+  
+  def cleaner(string)
+    @cleaners.each do |clean|
+      string = string.gsub(/#{clean}/i, ' ')
+    end
+    
+    [/(19|20\d{2})/, /\./, /\s*-\s*/, /\s{2,}/].each do |regex|
+      string = string.gsub(regex, ' ')
+    end
+    
+    string.strip
+  end
+  
+  def shortest(a,b)
+    # If the release contains a year, then the corresponding movie from IMDB should have the same year
+    if year = @search_value.match(/(19|20\d{2})/)
+      return nil unless year[1] == a[:year]
+    end
+     
+    Levenshtein.distance(self.super_cleaner(a[:title]), self.super_cleaner(self.cleaner(b)), @options[:limit])
+  end
+  
+  def super_cleaner(string)
+    string.gsub(/[^a-z0-9]/i, '')
+  end
 end

data/movie_searcher.gemspec

@@ -3,7 +3,7 @@ $:.push File.expand_path("../lib", __FILE__)
 
 Gem::Specification.new do |s|
   s.name        = "movie_searcher"
-  s.version     = "0.0.6"
+  s.version     = "0.0.8"
   s.platform    = Gem::Platform::RUBY
   s.authors     = ["Linus Oleander", "Jon Maddox"]
   s.email       = ["linus@oleander.nu", "jon@mustacheinc.com"]

data/spec/movie_searcher_spec.rb

@@ -17,10 +17,6 @@ describe MovieSearcher do
     MovieSearcher.find_by_release_name("asd asd asd asd asd asd asd asd asd asd asd asd asd asd asd asd asd asd asd asd asd asd asd").should be_nil
   end
   
-  it "should return nil when setting the limit to low" do
-    MovieSearcher.find_by_release_name('Paranormal Activity 2 2010 UNRATED DVDRip XviD-Larceny', :options => {:limit => 0}).should be_nil
-  end
-  
   it "should return the right movie" do
     [{
       :title => "Live Free Or Die Hard 2007 DVDRIP XviD-CRNTV", :iid => "tt0337978"
@@ -30,6 +26,10 @@ describe MovieSearcher do
       :title => "Heartbreaker 2010 LIMITED DVDRip XviD-SUBMERGE", :iid => "tt1465487"
     },{
       :title => "Paranormal Activity 2 2010 UNRATED DVDRip XviD-Larceny", :iid => "tt1536044"
+    }, {
+      :title => "The.Town.2010.EXTENDED.480p.BRRip.XviD-NYDIC", :iid => "tt0840361"
+    }, {
+      :title => "I Spit On Your Grave UNRATED 2010 DVDRip XviD-TWiZTED", :iid => "tt1242432"
     }].each do |movie|
       MovieSearcher.find_by_release_name(movie[:title]).imdb_id.should eq(movie[:iid])
     end
@@ -188,4 +188,50 @@ describe MovieSearcher, "should still have the same people" do
   it "should not have a role" do
     @movie.writers.first.role.should be_nil
   end
+end
+
+describe MovieSearcher, "should have a cleaner" do
+  before(:all) do
+    @movie = MovieSearcher.new({})
+  end
+  
+  it "should contain a list of at least 15 cleaners" do
+    @movie.should have_at_least(15).cleaners
+  end
+  
+  it "should clean up this release names" do
+    [{
+      :string => "Unstoppable 2010 BDRip XviD-REMALiA", 
+      :exclude => ["2010", "BDRip", "XviD"]
+    }, {
+      :string => "True Grit 2010 SCR XViD - IMAGiNE",
+      :exclude => ["2010", "SCR", "XViD", "IMAGiNE", "-"]
+    }, {
+      :string => "Black Swan 2010 DVDSCR XviD-ViSiON",
+      :exclude => ["2010", "DVDSCR", "XviD-ViSiON"]
+    }, {
+      :string => "Red 2010 720p BRRip XviD AC3-FLAWL3SS",
+      :exclude => ["720p", "BRRip", "XviD", "AC3-FLAWL3SS"]
+    }, {
+      :string => "I Spit On Your Grave UNRATED 2010 DVDRip 480p XviD-TWiZTED",
+      :exclude => ["UNRATED", "2010", "DVDRip", "XviD-TWiZTED", "XviD", "480p"]
+    }].each do |value|
+      string = @movie.cleaner(value[:string])
+      value[:exclude].each do |exclude|
+        string.should_not include(exclude)
+      end
+    end
+  end
+  
+  it "should not contain to many spaces in a row" do
+    @movie.cleaner("Unstoppable 2010 BDRip XviD-REMALiA").should_not match(/\s{2,}/)
+  end
+  
+  it "should not contain any dots" do
+    @movie.cleaner("The.Town.2010.EXTENDED.480p.BRRip.XviD-NYDIC").should_not match(/\./)
+  end
+  
+  it "should not contain any -" do
+    @movie.cleaner("The.Town.2010.EXTENDED.480p.BRRip.XviD-NYDIC").should_not match(/-/)
+  end
 end

metadata

@@ -1,13 +1,13 @@
 --- !ruby/object:Gem::Specification 
 name: movie_searcher
 version: !ruby/object:Gem::Version 
-  hash: 19
-  prerelease: 
+  hash: 15
+  prerelease: false
   segments: 
   - 0
   - 0
-  - 6
-  version: 0.0.6
+  - 8
+  version: 0.0.8
 platform: ruby
 authors: 
 - Linus Oleander
@@ -138,6 +138,7 @@ files:
 - README.md
 - Rakefile
 - lib/imdb_party.rb
+- lib/imdb_party/exclude.yaml
 - lib/imdb_party/httparty_icebox.rb
 - lib/imdb_party/imdb.rb
 - lib/imdb_party/levenshtein.rb
@@ -178,7 +179,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
 requirements: []
 
 rubyforge_project: 
-rubygems_version: 1.4.2
+rubygems_version: 1.3.7
 signing_key: 
 specification_version: 3
 summary: IMDB client using the IMDB API that their iPhone app uses